From the HBD Archive
Subject: Brewing Technique Statistics
Date: 1992-03-12 19:31:00 GMT

In #842, Steve Russell gives us a table regarding the 1990 national
competition, and writes:

>Besides, the jump from 67% to 84% in liquid yeast cited above
>is *tremendous* evidence that this is a great way to improve
>your beer. But again, it is only correlation, as statistical
>types will likely point out.

Unfortunately, it's not even that. I was going to put these data through
my stats program, but found the data weren't actually present in the
table. It was easy to convert the percentages in the "entries" column
to raw data, because the total number of entries wa given. But it's
not possible to convert the percentages in the "placed 1 2 or 3"
column to raw data, because the total number of such *winners* was not

Without knowing the size of the data set, I don't think it's possible
to determine if the difference in proportions is statistically
significant. CAVEAT: I am not a statistician, and I usually don't
analyze data in this form; I usually use ANOVA on data structures of
my own design.

Back New Search

The posts that comprise the Homebrew Digest Searchable Archive remain the property of their authors.
This search system is copyright © 2008 Scott Alfter; all rights reserved.